posTokens: POS Tagger

In myeomans/DTMtools: Adjustable Feature Counting for Text Analysis

Description

Tally parse-dependent features

Usage

posTokens(
  texts,
  ngrams = 1,
  language = "english",
  punct = FALSE,
  stop.words = TRUE,
  overlap = 1,
  sparse = 0.99,
  dependency = FALSE,
  tag.sub = 0,
  verbose = FALSE
)

Arguments

texts	a character vector of texts.
ngrams	numeric vector of ngram sizes (max = 1:3)
language	character what language are you parsing?
punct	logical should exclamation points and question marks be included as features?
stop.words	logical should stop words be included? default is TRUE
overlap	numeric How dissimilar (in cossine distance) must an ngram be from all (n-1)grams to be added to feature set?
sparse	maximum feature sparsity for inclusion (1 = include all features)
dependency	logical should features have dependency relations appended? default is FALSE
tag.sub	numeric what fraction of features should be replaced by POS tags? default is 0 (no features), fractions not supported yet.
verbose	logical - report interim steps during processing
POS	logical should features have part of speech tags appended? default is FALSE

Value

a matrix of feature counts

myeomans/DTMtools documentation built on March 2, 2020, 8:57 p.m.